Recurrent unexplained first-trimester miscarriage

(1)

Recurrent unexplained first-trimester miscarriage

Effects of acetylsalicylic acid, platelet aggregation and thyroid disease

Lennart Blomqvist

Department of Obstetrics and Gynecology Institute of Clinical Sciences

The Sahlgrenska Academy University of Gothenburg

Gothenburg, Sweden

(2)

Recurrent unexplained first-trimester miscarriage.

Effects of acetylsalicylic acid, platelet aggregation and thyroid disease

© 2019 Lennart Blomqvist blomqvist.lennart@telia.com

ISBN 978-91-7833-628-9 (PRINT) ISBN 978-91-7833-629-6 (PDF)

Printed by Brand Factory in Kållered, Sweden 2019

(3)

The most effective therapy for patients with unexplained RPL is often the most simple: antenatal counseling and psychological support.

Holly B Ford, 2009

(4)

(5)

Abstract

Background/Aims: Recurrent pregnancy loss (RPL) occurs in 1-2% of fertile couples and about 50% of cases are unexplained. Impaired placental circulation, increased platelet aggregation, immunological factors and thyroid autoimmunity have been suggested to be involved. Other placenta-mediated complications have been reduced by inhibition of platelet aggregation with acetylsalicylic acid (ASA).

The effect of ASA on RPL has been unclear. These studies aimed at investigating the effect of ASA treatment on RPL and arachidonic acid (AA)-induced platelet aggregation in women with RPL, as well as the effect of thyroid function by analyzing Thyroid Stimulating Hormone (TSH) and thyroid peroxidase antibodies (TPO-ab).

Methods: Women (n=640) with at least three unexplained first-trimester miscarriages were screened for inclusion in a single-center, randomized, placebo- controlled trial (the ASA-RCT, Paper I). Four hundred women were given either 75 mg ASA or placebo daily, beginning at gestational week (gw) 6-7, when fetal heartbeat was detected by vaginal ultrasound. Treatment ended at the latest at gw 36.

Treatment compliance was determined by analysis of AA-induced platelet aggregation using multiple electrode impedance aggregometry. All women underwent the same follow-up. Primary outcome was live birth rate (LBR).

In order to define reference values for the multiple electrode impedance aggregometry (the Multiplate analyzer), a longitudinal study was conducted including 79 healthy, non-smoking pregnant women with normal pregnancies (Paper II). Platelet aggregation induced by AA, adenosine diphosphate (ADP), thrombin receptor activating peptide 6 (TRAP) and collagen (COL) were determined four times during pregnancy and three months postpartum.

From the randomized population, 176 and 177 women, respectively, with normal

AA-induced platelet aggregation before pregnancy and treated with ASA or placebo,

were studied (Paper III). Platelet aggregation was determined before and during

pregnancy and results in the randomized groups were compared with one another, as

well as with those in the healthy controls from Paper II.

(6)

From the screened and eligible population, 495 women with complete thyroid test analyses [thyroid stimulating hormone (TSH), free thyroxine (T4) and thyroid peroxidase antibodies TPO-ab] before pregnancy were included. Risk factors for a new miscarriage were studied, in particular associations with TPO-ab and TSH in the upper normal range.

Results: In the ASA-RCT, all 400 randomized women completed the follow-up.

LBR were 83.0% and 85.5% in the ASA and placebo groups, respectively. The mean difference was -2.5% (95% CI to -10.1% to 5.1%). The risk ratio was 0.97 (95% CI 0.89 to 1.06).

In the longitudinal study of platelet aggregation during normal pregnancy, activation of platelets by AA, ADP and TRAP resulted in a minor decrease in platelet aggregation during pregnancy, compared with postpartum. COL-induced platelet aggregation was unchanged. A minor increase in platelet aggregation as pregnancy continued was found related to ADP.

There were no significant differences in AA-induced platelet aggregation when placebo-treated women with RPL were compared with healthy women with normal pregnancies. ASA treatment significantly reduced platelet aggregation during pregnancy, compared with before pregnancy. Gradually increased platelet aggregation was seen in the majority of ASA-treated women as pregnancy progressed. There were only two complete non-responders to ASA.

Miscarriage occurred more often in women with than without TPO-ab (25.7% vs 17.5%). There was no association between TSH in the upper normal range and a new miscarriage. Independent risk factors for a new miscarriage were age, number of previous miscarriages and presence of TPO-ab.

Conclusions: ASA does not prevent a new miscarriage in women with at least three previous first-trimester miscarriages. AA-induced platelet aggregation seems to be similar in women with RPL and in healthy women with normal pregnancies. ASA, 75 mg daily, decreases AA-induced platelet aggregation in most women during pregnancy, but the effect diminishes as pregnancy progresses. TPO-ab, but not TSH in the upper normal range, may be associated with an increased risk of a new miscarriage.

ISBN 978-91-7833-628-9 (PRINT)

ISBN 978-91-7833-629-6 (PDF)

(7)

Populärvetenskaplig sammanfattning

Bakgrund

Upprepade missfall drabbar omkring 1 på 100 av fertila par varav ungefär hälften inte har en påvisbar trolig orsak. Upprepade missfall i tidig graviditet kan definieras på olika sätt utifrån antal tidigare missfall och när i graviditeten de inträffat. I avhandlingens studier, har kvinnorna minst tre oförklarade missfall i följd under graviditetens första tredjedel och dessa tillsammans med samma partner.

Kvinnans ålder och antal tidigare missfall har i tidigare studier visats vara en riskfaktor för ett nytt missfall, men det finns också andra tänkbara förklarande orsaker; kromosomavvikelse hos kvinnan eller mannen, anatomiska avvikelser i livmodern, blodproppssjukdomar, hormonella sjukdomar (sköldkörtelsjukdom) samt livsstilsfaktorer (kraftig övervikt, extensiv träning). Dock, i fler än hälften av fallen visar utredning inte någon tänkbar eller påvisbar orsak.

Försämrad blodcirkulation med ”mikroblodproppar” i moderkakan och därmed sämre blodtillförsel till den nyanlagda graviditeten och fostret, har framförts som en orsak. Men även andra orsaker har föreslagits. Utan någon säkerställd effekt har olika behandlingar prövats såsom blodproppsförebyggande läkemedel (hämning av blodets levringsförmåga), sköldkörtelhormon (tyroxin), progesteron, immunglobulin.

Syftet med denna avhandling är att undersöka;

 om blodproppsförebyggande behandling med acetylsalicylsyra (ASA) kan minska risken för nytt missfall

 om funktionen hos blodplättar (trombocyter), små partiklar i blodet bidragande till blodets levringsförmåga, förändras under normal graviditet jämfört med kvinnors trombocytfunktion utan graviditet

 om trombocytfunktion under normal graviditet är annorlunda jämfört med graviditet föregången av upprepade tidiga missfall

 hur ASA behandling påverkar trombocyterna under graviditet

 om avvikelser i sköldkörtelfunktionen, närvaro av tyroperoxidas (TPO)-

antikroppar eller Tyroidea Stimulerande Hormon (TSH) i övre normal-

området, kan innebära ökad risk för ett nytt missfall

(8)

Arbete I

Denna studie gjordes för att utvärdera om behandling med 75 mg ASA dagligen, efter påvisad graviditet, kunde minska risken för ett nytt missfall hos kvinnor med minst tre tidigare upprepade missfall före tretton fulla graviditetsveckor och där utredning inte visat något avvikande. Paren remitterades till studien från kvinnokliniker på sjukhus samt öppenvård, hänvisades eller tog själva kontakt för att få delta i studien. De flesta kom från Region Västra Götaland och Region Halland.

Fyrahundra kvinnor, 200 kvinnor i varje grupp, lottades till ASA eller piller utan effekt, s.k. placebo, vid påvisad hjärtaktivitet hos fostret (graviditetsvecka 6-7) med ultraljud. Alla följdes med samma kontrollprogram under graviditeten.

Trombocytfunktionen mättes hos alla kvinnorna för att mäta effekten av ASA.

Resultatet visade att prognosen är mycket god för dessa kvinnor att föda barn oavsett ASA-behandling eller inte. Omkring 15% av kvinnorna fick ett nytt missfall inom varje grupp och 83% i ASA-gruppen och 85.5% i placebogruppen födde barn.

Följsamheten till behandlingen var mycket god.

Arbete II

Trombocyter och dess funktion under graviditet är av stor betydelse för vissa graviditetskomplikationer som preeklampsi (”havandeskapsförgiftning”), dålig fostertillväxt, för tidig födsel och kanske även för upprepade missfall. Det är viktigt att studera trombocytaggregationen (blodplättarnas ”hopklumpning”) under graviditet för att se om behandling med läkemedel som motverkar trombocyt- aggregationen kan minska risken för dessa graviditetskomplikationer.

Sjuttionio av 104 friska gravida kvinnor utan tidigare upprepade missfall och utan komplikationer i aktuell graviditet, studerades avseende trombocytfunktion under graviditet och tre månader efter förlossning då trombocytfunktionen bedöms ha återgått till en normal nivå.

Resultatet visade en något minskad trombocytaggregation under graviditet jämfört med tre månader efter förlossningen.

Arbete III

Arbete III gjordes för att jämföra trombocytaggregationen under graviditet hos

kvinnor med upprepade missfall med dem som haft ett eller inget missfall, samt hur

trombocytaggregationen påverkades av behandling med 75 mg ASA respektive med

placebo. Trombocytaggregation studerades dessutom för att få kunskap om ASA-

(9)

effekten vid olika tidpunkter under graviditeten och jämförelse gjordes med trombocytaggregation före behandling.

Resultaten visade att, det inte var någon skillnad i trombocytaggregation som aktiverats av arachidonsyra (AA) under graviditet, för kvinnor med upprepade missfall och de normala kontrollerna. Det finns troligen inget stöd för hypotesen att kvinnor med upprepade missfall skulle ha en ökad AA-aktiverad trombocytaggregation som förklaring till missfallen.

Trombocytaggregationen hos de ASA-behandlade kvinnorna visade mer än 50%

reduktion, vilken kvarstod för de allra flesta under hela graviditeten. Reduktionen minskade dock något under graviditetens senare del. Detta är inte visat tidigare och skulle kunna visa på ett behov att under senare del av graviditeten höja dosen av ASA om 75 mg per dygn används som prevention mot de tidigare nämnda graviditets- komplikationerna.

Arbete IV

Underfunktion av sköldkörteln (hypotyreos) är en känd riskfaktor för missfall. Av den anledningen kontrolleras körtelns funktion med ett blodprov, TSH, vid inskrivning på mödravården i tidig graviditet, så att behandling med tyroxin kan sättas in vid behov. Ett annat tecken till påverkan på sköldkörteln är förekomst av TPO-antikroppar i blodet. Det saknas idag kunskap om TSH i sitt övre normalområde (2.5 <TSH ≤4.0 mU/l) med eller utan samtidig förekomst av TPO-antikroppar är riskfaktorer för nytt missfall hos kvinnor med tidigare upprepade missfall i första tredjedelen av graviditeten.

I detta arbete ingick 495 av 640 kvinnor som tidigare utretts för att kunna ingå i studien i arbete I. Alla hade normala prover avseende TSH och fritt T4.

Graviditetsutfall (missfall och födsel) jämfördes avseende förekomsten av TPO- antikroppar samt nivån av TSH.

Resultaten visade att TSH i sitt övre normalintervall inte var en riskfaktor för nytt

missfall, men att förekomst av TPO-antikroppar skulle kunna vara det.

(10)

(11)

List of papers

I Blomqvist L, Hellgren M, Strandell A.

Acetylsalicylic acid does not prevent first-trimester unexplained recurrent pregnancy loss: A randomized controlled trial. Acta Obstet Gynecol Scand.

2018;97:1365-1372.

II Blomqvist L, Strandell A, Baghaei F, Hellgren M.

Platelet aggregation in healthy women during normal pregnancy - a longitudinal study. Platelets. 2019;30:438-444.

III Blomqvist L, Strandell A, Jeppsson A, Hellgren M.

Platelet aggregation during pregnancy in women with previous recurrent first- trimester fetal loss, with and without acetylsalicylic acid treatment. Submitted.

IV Blomqvist L, Filipsson Nyström H, Hellgren M, Strandell A.

Preconceptual thyroid peroxidase antibody positivity in women with recurrent

pregnancy losses may be a risk factor for another miscarriage. In manuscript.

(12)

(13)

List of abbreviations

AA Arachidonic acid

ACOG American College of Obstetricians and Gynecologists

ADP Adenosine diphosphate

APL Apoteket Produktion & Laboratorier APTT Activated partial thromboplastin time APS Antiphospholipid syndrome

ASA Acetylsalicylic acid

ASPI-test Reagent kit used for monitoring of ASPIRIN and other

cyclooxygenase inhibitors, after induction by arachidonic acid ASRM American Society for Reproductive Medicine

AUC Area under curve

BMI Body mass index

BV Bacterial vaginosis

CI Confidence interval

COL Collagen

COX Cyclooxygenase

DGGG German Society of Gynecology and Obstetrics (Deutsche Gesellschaft für Gynekologie und Geburtshilfe)

ESHRE European Society of Human Reproduction and Embryology Free T4 Free thyroxine

gw Gestational week

Hb Hemoglobin

hCG Human chorionic gonadotropin

ITAKA Identify of TAblets and KApslar (capsulas) IVF In vitro fertilization

IVIG Intravenous immunoglobulin

LBR Live birth rate

(16)

LC Leucocyte count

LMWH Low molecular weight heparin NK Natural killer cell

OEGGG Austrian Society of Gynecology and Obstetrics (Oesterreiche Gesellschaft für Gynekolgie and Geburtshilfe)

OR Odds ratio

PC Platelet count

PC(INR) Prothrombin complex (international normalized ratio) PCOS Polycystic ovary syndrome

PFA-100 Platelet Function Analyzer PI Principal investigator PRP Platelet rich plasma

RCOG Royal College of Obstetricians and Gynecologists RCT Randomized controlled trial

REPL Recurrent early pregnancy loss RPL Recurrent pregnancy loss

RR Risk ratio

SCH Subclinical hypothyroidism

SD Standard deviation

SFOG Swedish Society of Obstetrics and Gynecology

SGGG Swiss Society of Gynecology and Obstetrics (Schweizerische Gesellschaft für Gynekologie und Geburtshilfe)

TLC Tender loving care

TPO-ab Thyroid peroxidase antibody

TRAP Thrombin receptor activated peptide 6 TSH Thyroid Stimulating Hormone

TXA

2

Thromboxane A2

U Unit

UK United Kingdom

WHO World Health Organization

(17)

17

Introduction

Miscarriage

Miscarriage, the spontaneous loss of a pregnancy before the fetus has reached viability (Rai et al., 2006), is the most common complication of pregnancy.

Miscarriage, also called spontaneous abortion, can be classified as sporadic or recurrent. A sporadic miscarriage occurs in 30-50% of pregnancies, but only 15% of cases are diagnosed (van Dijk et al., 2016; Rai et al., 2006; Kutteh et al., 2014).

Between 12% and 24% of clinically diagnosed pregnancies end in miscarriage, usually in the first trimester (Lee et al., 1996), as illustrated in Figure 1 (Nybo- Andersen, 2000).

Fig. 1. Weekly risk of fetal loss. Anne-Marie Nybo Andersen. Fetal death, epidemiological studies. PhD thesis. University of Copenhagen. 2000.

Published with permission from the author.

Age and risk of miscarriage

Many authors have reported an increased risk of fetal death and spontaneous abortion with increasing maternal age, especially over 30 years, irrespective of previous obstetric history (Nybo-Andersen et al., 2000; Cauchi et al., 1991; Lund et al., 2012;

Bhattacharya et al., 2010).

Reproductive behavior in our modern society has changed; many women now choose

to delay pregnancy and childbearing. According to UK data, the number of babies

born to women aged 35 years and up doubled between 1985 and 2001, from 8% to

(18)

18

16% of all births (Rai et al., 2006). In 1973, women in Sweden had their first child at a mean age of 24 years. In 2016, the corresponding age was 29.2 and in 2018 it was 29.4. In Danderyd, a municipality near Stockholm, the mean age at first childbirth was 31.8 years (highest in the country) in 2018, while it was 31.6 in Stockholm, 30.3 in Gothenburg (the second biggest city in Sweden) and 28.3 in Borås (a middle-sized city, in which the main study was conducted) (Statistics Sweden, 2018). In smaller cities and in rural areas, women had their first child at age 25-26. The level of education and women wanting to establish themselves on the labor market before considering pregnancy and starting a family have been suggested as factors contributing to this variation.

Nybo-Andersen conducted a prospective register study on data from the Civil Registration System in Denmark (Nybo-Andersen et al., 2000) between 1978 and 1992, including 634,272 women and the outcome of 1,221,546 pregnancies. The overall risk of fetal loss, defined as spontaneous abortion, hydatiform mole, ectopic pregnancy or stillbirth, was 13.5%. More than 20% of all pregnancies in 35-year-old women ended in fetal loss, while this occurred in 54.5% of women aged 42 years.

About 80% of fetal losses were spontaneous abortions. The risk of spontaneous abortion increased with age: 8.7% in women aged 22 years and 84.1% in women aged 48 and up (Fig. 2) and number of preceding spontaneous abortions (Fig. 3) (Nybo-Andersen et al., 2000).

Fig. 2. Risk of spontaneous abortion according to maternal age at conception, stratified according to calendar period (Nybo-Andersen et al. Maternal age and fetal loss: population- based register linkage study. BMJ 2000;320:1708).

Published with permission fromBMJ/Copyright Clearance Center.

(19)

19

Fig. 3. Risk of spontaneous abortion in nulliparous women (left panel) and parous women (right panel), according to maternal age at conception and number of spontaneous abortions in preceding 10 years (Nybo-Andersen et al. Maternal age and fetal loss: population-based register linkage study BMJ 2000;320:1708). Published with permission from BMJ/Copyright Clearance Center.

The study demonstrated a clear independent effect of maternal age on the risk of spontaneous abortion, in addition to an increased risk of ectopic pregnancy and stillbirth with increasing maternal age (Nybo-Andersen et al., 2000).

Recurrent pregnancy loss

Empirically, recurrent pregnancy loss (RPL) occurs in 1-2% of fertile couples, if defined as at least three consecutive losses, and in 5% if the definition is two or more losses. According to the literature, RPL affects 0.5% to 2.3% of couples if the definition is at least three losses and 5% if it is defined as at least two losses (Stirrat, 1990; Jauniaux et al., 2006; Rasmark Roepke et al., 2017). If only confirmed (by ultrasound and/or histology) pregnancies were included, an incidence of 0.8 to 1.4%

has been reported. Adding loss of biochemically diagnosed pregnancies increased the incidence to 2-3% (Larsen et al., 2013).

The definition of RPL is thus debated, ranging from two, not necessarily consecutive, to three consecutive miscarriages. Unfortunately, the terminology in the literature is not consistent. It is difficult to compare results of different studies due to this lack of consensus regarding nomenclature and classification of RPL (Kolte et al., 2015).

Most authors seem to have adopted the definitions at least three consecutive pregnancy losses before gestational week (gw) 13 (Clifford et al., 1997; Rai et al., 2000; Wilson et al., 1999) or gw 20 (Jablonowska et al., 1999; Abu-Heija, 2014;

Kaur et al., 2016). In recent years, many authors have adopted the definitions at least two losses before gw 13 (D´Uva et al., 2008; Bernardi et al., 2013; Pasquier et al.,

Risk of spontaneous abortion (%)

≤19 20-24 25-29 30-34 35-39 40-44 ^≥45

Maternal age at conception

Risk of spontaneous abortion (%)

≤19 20-24 25-29 30-34 35-39 40-44 ≥45 Maternal age at conception

(20)

20

2015) or gw 20 (Kaandorp et al., 2010; Vissenberg et al., 2015; van Dijk et al., 2016).

Some authors apply gw 24 in their definition instead (Brigham et al., 1999; Clark et al., 2010). Different guidelines are presented in Table 1.

Table 1. Definitions of recurrent pregnancy loss in international guidelines

Guideline Definition

American College of Obstetricians and Gynecologists (ACOG), 2016

Two or more pregnancy losses before gw 24

American Society for Reproductive Medicine (ASRM), RPL Guidelines 2013

Two or more consecutive miscarriages

European Society of Human Reproduction and Embryology (ESHRE), RPL Guidelines Nov. 2017;

Hum Reprod Open 2018

Two or more pregnancy losses before gw 24 (RPL), or before gw 10 (REPL)

Royal College of Obstetricians and Gynecologists (RCOG), Green-top Guideline No.17, 2011

Three or more consecutive pregnancy losses before gw 22

Swedish Society of Obstetrics and Gynecology (SFOG), Guidelines March 2017

Three or more consecutive miscarriages before gw 22

World Health Organization (WHO), 1976 Three or more consecutive miscarriages before gw 20

German (DGGG), Austrian (OEGGG) and Swiss (SGGG) Society of Gynecology and Obstetrics;

Geburtshilfe und Frauenheilkunde 2018;78:364-381

Three or more consecutive miscarriages before gw 20

RPL=recurrent pregnancy loss, REPL=recurrent early pregnancy loss, gw=gestational week

RPL is classified as either primary, with no previous live births, or secondary, in which recurrent miscarriages are preceded by live births.

About 50% of RPL are unexplained (Stirrat, 1990; Quenby et al., 1993; Clifford et al., 1997; Yadava et al., 2014; Homer, 2019).

Based on data from the Danish National Patient Register, Knudsen et al. conducted a study on a large unselected Danish population consisting of 300,500 pregnancies, of which about 33,900 were spontaneous abortions (Knudsen et al., 1991). The objective was to investigate the overall risk of a new miscarriage, defined as a spontaneous abortion before gestational week 28, related to the number of previous miscarriages (Table 2). Their results are compared to results reported by other authors during different time periods in Table 3.

To evaluate the live birth rate (LBR) within five years of the RPL investigation, 987

women with at least three consecutive miscarriages were monitored in a prospective

register study by Lund. The LBR was 81.3% for women aged 20-24 years, 69.9%

(21)

21

for women aged 30-34 years and 41.7% for women over age 40. Moreover, the number of previous miscarriages had an effect on the LBR, which was 71.9% after three and 50.2% after six miscarriages (Lund et al., 2012) (Fig. 4).

Table 2. Risk of a new miscarriage according to age and

number of previous miscarriages

Number of previous abortions

Age <35 years risk (%)

Age ≥35 years risk (%)

0 9.7 24.2

1 14.9 24.0

2 23.8 32.8

3 43.0 54.8

4 53.8 56.3

Modified table from Knudsen et al., 1991 (Eur J Obstet Gynecol Reprod Biol 1991;39:31).

Table 3. Risk of spontaneous abortion expressed in percentage, according to

number of previous consecutive spontaneous abortions

Author, Year Number of previous miscarriages

1 2 3 4

Whitehouse, 1930 22 38 73 -

Eastman, 1946 13 37 84 -

Warburton, 1964 24 26 32 -

Knudsen, 1991 16 25 45 54

Fig. 4. Chance of a live birth after consultation for recurrent pregnancy loss by number of previous miscarriages (Lund et al. Recurrent Miscarriage and Prognosis for Live Birth. Obstet Gynecol 2012;119:37).

Published with permission from Lippincott Williams & Wilkins.

(22)

22

Conceivable and probable causes of recurrent pregnancy loss

Some causes of early RPL are regarded as treatable. There are etiological factors, possible to identify and sometimes to treat, in up to 50% of RPL (Clifford et al., 1997; Bricker et al., 2002; Kutteh et al., 2014). The most important of these are chromosomal aberration, reproductive tract malformation, luteal phase insufficiency, antiphospholipid syndrome, genital infection, hypercoagulable state, iatrogenic factors and lifestyle factors.

Chromosomal aberration

Impaired fertility and increased risk of first-trimester RPL were found when an aberration was carried by either the man or the woman. On investigation of RPL, a chromosomal aberration prevalence of 2-5% was found, mostly consisting of balanced translocations, compared with 0.7% in the general population (Homer 2019; Ford et al., 2009; Rai et al., 2006). Two types of balanced translocation have been reported: reciprocal balanced translocation, in which there is an exchange of two terminal segments from different chromosomes, and Robertsonian translocation, i.e. centric fusion of two acrocentric chromosomes. The most frequent versions are between chromosomes 13 and 14, 14 and 15 and 14 and 21, respectively. In the female, this can present as 45,XX,rob(14;15)(q10;q10). The carrier of a balanced translocation has a normal phenotype and no sign of chromosomal injury but can produce unbalanced gametes and an unbalanced fetus. In general, the translocation prevalence is higher in females than in males (Venkateshwari et al., 2010).

Reproductive tract malformation

Congenital uterine anomalies are found in 1.8% to 37.6% of women with RPL (Rai et al., 2006; Ford et al., 2009; Morley et al., 2013). Intrauterine septum is the most common congenital anomaly, and that which is most closely linked to RPL. Other important uterine anomalies are fibroids and polyps. Depending on the size and location of the anomaly, it can create poorly vascularized endometrium for implantation, and is thus a plausible reason for the miscarriage.

Luteal phase insufficiency with progesterone deficiency

Women with polycystic ovary syndrome (PCOS) are reported to have a RPL prevalence of 8-20% (Bricker et al., 2002; Kutteh et al., 2014; Rai et al., 2000).

Insufficient progesterone exposure during early pregnancy results in impaired

secretory endometrium, which is important for normal embryo implantation and

(23)

23

growth (Palomba et al., 2015). A positive effect of vaginal progesterone treatment initiated in the luteal phase was demonstrated in a randomized placebo-controlled trial of 700 women with unexplained RPL (Ismail et al., 2018). Highly significant differences in miscarriage rates (12.4% vs. 23.3%) and LBR rates (91.6% vs. 77.4%) were reported. However, a systematic review and meta-analysis, including the above-mentioned randomized controlled trial (RCT), concluded that there were too few studies to draw firm conclusions about the efficacy of progesterone treatment for women with RPL without specifically addressing luteal phase insufficiency (Rasmark Roepke et al., 2018).

Antiphospholipid syndrome

Antiphospholipid syndrome (APS) is an autoimmune disease that interferes with coagulation (Yadava et al., 2014). The prevalence of APS in women with RPL is 5- 20% (Bricker et al., 2002; Rai et al., 2006; Kutteh et al., 2014; Yadava et al., 2014;

Homer, 2019), compared with 2-5% in unselected obstetric patients (Branch et al., 2010). APS is the most important treatable cause of RPL. It is diagnosed by moderate to high blood titers of lupus anticoagulant, anti-cardiolipin antibodies and/or anti-β2- glucoprotein-1 antibodies, at two testing occasions with an interval of at least 12 weeks (Homer et al., 2019). It is a major risk factor underlying adverse pregnancy outcomes, including RPL (Kutteh et al., 2014), and the only thrombophilia with a proven association to RPL (McNamee et al., 2012). Women with the syndrome have a miscarriage rate of 90% in subsequent untreated pregnancies (Rai et al., 2006). The mechanisms by which APS results in RPL are insufficiently understood (Ford et al., 2009) but thrombosis and infarction of the uteroplacental vasculature have been suggested as mediators (Sebire et al., 2002).

Infection

Bacterial vaginosis (BV) in the first trimester is associated with an increased risk of

late miscarriage or preterm birth although causality has not been proven (Larsson, et

al., 2005). The prevalence of BV in pregnant women is 15.6% to 32.5% (Svare, et

al., 2006; Jacobsson, et al., 2002; McGregor, et al., 1995). BV was more frequent in

women with a history of late miscarriage (21%) than in women with recurrent

pregnancy loss (8%) (Llahi-Camp, et al., 1996), and there was no significant

association between BV and a history of RPL (Gözde, et al., 2016). The value of

screening and treatment in early pregnancy is under debate.

(24)

24

Inherited thrombophilia

Prospective cohort studies have not shown any significant association between RPL and thrombophilia, including factor V Leiden mutation, prothrombin gene mutation, protein C deficiency, protein S deficiency and antithrombin deficiency.

Iatrogenic and lifestyle factors

Smoking, consumption of alcohol and caffeine intake are not associated with the risk of RPL (Zhang et al., 2010). Couples with RPL should, however, be informed that smoking and alcohol could have a negative impact on their chances of live birth, and cessation is thus recommended nonetheless (ESHRE Nov., 2017). Maternal obesity (BMI ≥30 kg/m

²⁾

significantly increased the risk of miscarriage in couples with unexplained RPL. There was no increased risk in non-obese women with overweight (Lo et al., 2012). No studies were found regarding the impact of exercise on the chances of live birth in women with RPL. The effect of exercise on the risk of a sporadic first-trimester miscarriage is unclear. In a review by Hegaard (2016) two studies reported that exercise was associated with a lower risk of miscarriage (Xu et al., 2014; Zhang et al., 2011), two others showed the same risk of miscarriage in exercising and non-exercising pregnant women (Maconochie et al., 2007; Clapp, 1989) and one found an increased risk of first-trimester miscarriage in exercising women (Madsen et al., 2007). High-intensity occupational activity has been identified as a risk factor for miscarriage (Schlussel et al., 2008). Treatment for cervical intraepithelial neoplasia (conization) was not associated with early miscarriage and RPL, although it was associated with a significantly increased risk of second-trimester miscarriage (Kyrgiou et al., 2014).

Unexplained recurrent pregnancy loss

As mentioned above, about 50% of RPL are unexplained. This thesis focuses on investigating the respective roles of acetylsalicylic acid (ASA), platelet aggregation and thyroid function.

Platelet aggregation and hemostasis

Impaired platelet aggregation has been suggested to play an important role in

placenta-mediated obstetric complications, such as RPL, preeclampsia, intrauterine

growth retardation, impaired placental circulation and pregnancy-induced

hypertension (Morrison et al., 1985; Dogan Gun et al., 2006; Bujold et al., 2009;

(25)

25

Flood et al., 2010; Sultana et al., 2012; Dempsey et al., 2015; Atallah et al., 2017).

The role of platelet aggregation in unexplained recurrent miscarriage is unclear. No longitudinal study of platelet aggregation in women with at least three consecutive unexplained first-trimester RPL has been found. Furthermore, no published trial has demonstrated the effect of low-dose ASA on platelet aggregation during pregnancy in this group.

Platelet aggregation

As platelet reactivity is an important factor in placenta-mediated obstetric complica- tions, it is of great interest to study platelet aggregation during pregnancy, since increased aggregation might be a factor associated with these complications. There are different platelet aggregation receptors on the platelet´s surface (Fig. 5).

Platelet aggregation can be induced by different activators and analyzed with different methods, such as light transmission aggregometry, multiple electrode impedance aggregometry and platelet function analyzer (PFA-100) (Zimmermann et al., 2008; Lordkipanidzé et al., 2007). Born light transmission aggregometry is considered to be the gold standard (Le Quellec et al., 2016).

Fig. 5. Receptors on the platelet´s surface through which different activators induce platelet aggregation. Published with permission from Roche Diagnostics.

Four activators were used in these studies:

Arachidonic acid (AA) is often used for monitoring ASA and other cyclooxygenase

(COX) inhibitors. Platelet aggregation induced by AA is the substrate for COX. COX

forms thromboxane A2 (TXA

2

) which is a potent platelet agonist and vaso-

constrictor.

(26)

26

Collagen (COL) leads to platelet activation via the platelet COL receptor, which generates release of AA, which is converted to TXA

2

.

Adenosine diphosphate (ADP) activates platelet aggregation through several receptors on the platelet surface, the most important of which is P2Y12.

Thrombin receptor activating peptide 6 (TRAP) stimulates platelet aggregation via the thrombin receptor. Thrombin is a very potent platelet activator.

Multiple electrode impedance aggregometry (Multiplate analyzing system) was used for platelet aggregation analysis in our studies. This method was chosen because of many advantages: the use of whole blood, no time-consuming centrifugation is needed, results are available within a few minutes, only a very small amount of blood is needed and the method is easy to handle.

The analysis takes place in a single-use test cell containing a dual sensor unit and a Teflon-coated stirring magnet. Upon activation, the platelets adhere to and aggregate on metal sensor wires in the Multiplate test cell. Every Multiplate test cell contains two independent sensor units, each consisting of two silver-coated, highly conductive copper wires with a length of 3.2 mm. When platelets are activated they attach to vascular injuries as well as to artificial surfaces. When they adhere to the sensor wires the electrical resistance between the wires rises (Fig. 6). The increase in impedance is transformed into aggregation units (U) and plotted against time. Two curves are assessed using the two independent sensors in the test cell. The parameters calculated by the software are the mean values of the parameters determined with each curve.

The result is shown as the area under the curve (AUC) and presented in units (U) (Fig. 7).

Fig. 6. Activated platelets adhere to the metal sensor wires in the Multiplate test cell and the electrical resistance between the wires rises. The increase of impedance is converted into aggregation units and plotted against time. Published with permission from Roche Diagnostics.

(27)

27 Fig. 7.

Area

Fig. 7. Area under the curve denotes platelet aggregation after activation by arachidonic acid, without medication (left panel) and under treatment with ASA (right panel). Red and blue denote parallel tests. Published with permission from Roche Diagnostics.

U=aggregation units

Multiplate multiple electrode impedance aggregometry has been used in the evaluation of platelet inhibitors during treatment of non-pregnant individuals (Jambor et al., 2009; Mueller et al., 2009; Paniccia et al., 2009; Le Quelles et al., 2016). When it comes to pregnancy, the method has been used in a small study comparing platelet aggregation during normal pregnancy and preeclampsia, as well as in a larger study concerning non-responsiveness to ASA for prevention of preeclampsia (Can et al., 2010; Navaratnam et al., 2018).

Normal pregnancy

Normal pregnancy leads to changes in the coagulation and fibrinolytic systems, i.e.

increased activity of a number of clotting factors, decrease in protein S levels, inhibition of fibrinolysis, and reduced activity of activated protein C, an important anticoagulant (Bremme, 2003).

Platelet reactivity during normal pregnancy remains poorly characterized (Valera et al., 2010). No longitudinal study of platelet aggregation induced by different agonists in healthy women with uncomplicated pregnancy has been found.

Conflicting results have been published regarding platelet aggregation during pregnancy, perhaps due to the use of different agonists and various anticoagulants used in blood sampling.

200 U

0 U

(28)

28

Norris conducted a small study (n=20) investigating platelet aggregation in healthy primigravidas at gw 12, 20, 28, 32 and 36; during labor; at 1, 24 and 48 hours after delivery; and six weeks postpartum (Norris et al., 1993). Platelet aggregation was induced by AA, COL and ADP and measured by a modification of the method described by Fox (1982). Pregnancy had a significant effect on COL-induced platelet aggregation from gw 20, reaching a maximum during labor but decreasing sharply within 48 hours after delivery. Only a small, non-significant increase remained at six weeks postpartum. Platelet aggregation induced by ADP exhibited a significant increase at gw 36 and a slight increase at six weeks after delivery. Pregnancy had a significant effect on AA-induced platelet aggregation at gw 32 and 36, compared with gw 20, peaking during labor and gradually decreasing until 48 hours after delivery. In summary, the findings of this study were that platelet aggregation induced by AA and COL increased during normal pregnancy. Platelet aggregation was inhibited by ASA (Norris et al., 1993).

Obstetric complications

Platelet aggregation induced by AA and COL has been studied in pregnant women with gestational hypertension, essential hypertension and preeclampsia (Morrison et al., 1985; Norris et al., 1993).

In a study by Whigham, aimed at evaluating platelet aggregation in preeclampsia, two control groups were included, non-pregnant women (n=11) and women with a normal pregnancy (n=9). No significant difference was found between the two control groups in platelet aggregation induced by AA, ADP and COL in a single sample drawn at gw 36. Platelet aggregation was measured photometrically in 0.1 ml of platelet-rich plasma (PRP), using the Bryston aggregometer described by Gordon and Drummond (1974) (Whigham et al., 1978).

Recurrent pregnancy loss

In a study by Flood, platelet aggregation in 30 women with at least three consecutive

miscarriages before gw 20, and in 30 healthy women with normal pregnancy and no

adverse obstetric history (one or no previous miscarriage), was investigated (Flood

et al., 2010). Single samples were drawn in the non-pregnant state at least 12 weeks

after the last miscarriage. There was no significant difference between the RPL and

control groups in platelet aggregation induced by ADP, COL and TRAP. Women

with unexplained RPL had significantly increased platelet aggregation in response to

AA (Flood et al., 2010).

(29)

29

A study was conducted by Dempsey in order to evaluate platelet function in very early pregnancy in patients with at least three consecutive unexplained miscarriages.

Patients were divided into two cohorts: women with a viable pregnancy at gw 12 and a subsequent live birth (n=30) and women who miscarried before gw 12 (n=9). A single blood test was drawn at gw 4-7. Platelet aggregation was analyzed with a modification of light transmission aggregometry, after induction by AA, ADP, COL and TRAP. In women with unexplained RPL, in whom pregnancy ended in a new miscarriage, there was a significant reduction in platelet aggregation in response to ADP and TRAP, but not to AA or COL (Dempsey et al., 2015).

Thyroid function

General aspects

The effect of thyroid disease and autoimmunity on adverse pregnancy outcome, such as decreased LBR, premature delivery, miscarriage and first-trimester RPL, has been pointed out by many authors (van den Boogaard et al., 2011; Thangaratinam et al., 2011; Mannisto et al., 2013; Dhanwai Kumar et al., 2016), but is also under debate.

Hypothyroidism and subclinical hypothyroidism (SCH), defined as elevated serum thyroid stimulating hormone (TSH) with normal serum free thyroxine (free T4) levels (Vissenberg et al., 2012; van Dijk et al., 2016), are regarded as causes of adverse pregnancy outcome (Chan et al., 2015; Faisal et al., 2016). The effect of SCH in women with miscarriage is unclear. Some authors report an increased risk of miscarriage (van den Boogaard et al., 2011; Liu et al., 2014; Maraka et al., 2016; Ma et al., 2016), while some found no association at all (Bernardi et al., 2013; van Dijk et al., 2016; Plowden et al., 2016). Thyroid autoimmunity, detected as thyroid antibodies, mostly thyroglobulin antibodies and thyroid peroxidase antibodies (TPO- ab), has been associated with increased risk of sporadic and recurrent miscarriage (Negro et al., 2006; Vissenberg et al., 2012; Lata et al., 2013; Meena et al., 2016;

Wang et al., 2017).

Thyroid peroxidase is an enzyme expressed in the inner membrane of the follicular

cell within the thyroid gland (Ruf et al., 2006) and is normally undetectable by the

immune system. The production of TPO-ab is regarded as a marker of autoimmune

thyroid disease, as the integrity of the follicular cell is disturbed and will put the

patient at risk of future thyroid dysfunction (Sadler et al., 2012).

(30)

30

Several pathophysiological theories have been presented to explain the proposed association between thyroid hormone disturbances, with or without TPO-ab, and subfertility and early pregnancy loss, but the exact mechanism is unknown (van den Boogaard et al., 2011). Some theories have focused on endometrial changes, involving T cells and cytotoxic natural killer (NK) cells (Nazarpour et al., 2017), diminished expression of trophoblast endocrine function (Maraka et al., 2017), down-regulation of the thyroid hormone receptors THR-A and THR-B in villous trophoblasts (Ziegelmüller et al., 2015) or endometrial physiology in the implantation process (Aghajanova et al., 2011). More general theories include autoimmunity against the fetal allograft (Prummel et al., 2004) and thyroid hormone deficiency during implantation and in the early stages of embryo development (Coliccia et al., 2014; Vissenberg et al., 2015).

Recurrent pregnancy loss

The prevalence of TPO-ab in euthyroid women of fertile age is 8-14% (Yan et al., 2012; Lin et al., 2014; Vissenberg et al., 2015). In euthyroid women with RPL, defined as at least three unexplained first-trimester miscarriages, positive TPO-ab were found in 10.7% and 23.5%, respectively, in two controlled studies (Ticconi et al., 2011; Yan et al., 2012).

A two- to four-fold increased risk of both sporadic and recurrent miscarriage has been reported by many investigators when TPO-ab are detected in euthyroid women (Negro et al., 2005; Ticconi et al., 2011; Thangaratinam et al., 2011). However, other authors have not found any such association (Esplin et al., 1998; Lata et al., 2013;

Plowden et al., 2016).

The ultimate proof of thyroid involvement in RPL would be if thyroxine treatment reduced the risk of a new miscarriage. Again, some studies, both observational and randomized, found an effect of treatment (Negro et al., 2006; Glinoer et al., 2006;

Lapoutre et al., 2012; Nazarpour et al., 2017), while other RCTs found no positive

effect (Negro et al., 2016; Wang et al., 2017). The TABLET trial was published in

2019, reporting that levothyroxine treatment before conception, in euthyroid women

with one or more miscarriages and TPO-ab, neither decreased the miscarriage rate

nor increased the LBR. The fact that women with as few as one miscarriage were

included, and that no common laboratory threshold for TPO-ab levels was applied to

all participating centers, were limitations (Dhillon-Smith et al., 2019).

(31)

31

The European Society of Human Reproduction and Embryology (ESHRE) concluded in their 2017 guideline that the evidence regarding treatment with levothyroxine in women with SCH and RPL is conflicting. Further investigation of this potential treatment effect is thus required. Regarding positive TPO-ab and levothyroxine treatment, the published studies are too small to draw any robust conclusion. There is insufficient evidence to support treatment with levothyroxine in euthyroid women with TPO-ab and RPL outside a clinical trial (ESHRE Guidelines Nov., 2017).

One problem, when comparing studies, is the different definitions of positive TPO- ab, which varies from 10 to 200 kU/l, according to the respective laboratory.

Furthermore, the TSH reference values for definition of SCH vary. However, the lower cut-off level is 2.5 mU/l in the majority of the studies.

Suggested treatments for unexplained recurrent pregnancy loss

Different treatments, based on different theories regarding plausible etiologies, have been attempted for unexplained RPL. ASA, low-molecular-weight heparin (LMWH), intravenous immunoglobulin (IVIG), corticosteroids, lipid emulsion and leucocyte immunization have all been tried. There is no conclusive evidence that any currently available specific medical intervention is successful in decreasing the risk of miscarriage in women with unexplained RPL (Dempsey et al., 2015; Rasmark Roepke et al., 2018).

Pharmacological treatment with ASA, LMWH, IVIG and leucocyte immunization have been evaluated in RCTs. For ASA and LMWH summarized in Table 4. The concept of “tender loving care” has also been studied, albeit only in observational studies.

Antiplatelet and anticoagulant therapy

The historical hypothesis behind RPL was that affected women are already in a pro- thrombotic state before pregnancy; indeed, a positive effect of ASA in prevention of placental thromboses has been reported (Jahaninejad, 2014).

Anti-platelet or anticoagulant drugs, or a combination, have been suggested in order

to prevent subsequent miscarriage in women with previous early RPL. In addition to

(32)

32

preventing thrombosis, these suggestions have been grounded in different hypotheses concerning disturbed placental circulation, a pro-thrombotic female phenotype or coagulation-cascade activation of the developing trophoblasts (Geer et al., 2010).

Pregnancy is a hypercoagulable state with increased levels of pro-coagulant factors (Stirling et al., 1984) and increased number of circulating endothelial microparticles (Carp et al., 2004). It has thus been suggested that miscarriage is a coagulopathy (Rai, 2003). Significantly increased platelet aggregation induced by AA has been reported in women with unexplained RPL (Flood et al., 2010). Furthermore, this theory is supported by the fact that microthrombi and thromboses are common in placental vessels (Rushton, 1988; Dogan-Gun et al., 2006).

Acetylsalicylic acid

A precursor to ASA, found in willow leaves, has been used for health effects for at least 2,400 years. In 1853, the chemist Charles Frédéric Gerhard produced ASA for the first time. During the following fifty years, the chemical structure was established and more efficient production methods were developed. In 1899, Bayer named it aspirin, and its popularity spread around the world. Today, ASA is one of the most widely used medications globally, with about 44,000 tons, or 100 billion pills, consumed each year (Jones et al., 2015). It is included in the World Health Organization´s (WHO) List of Essential Medicines that includes the safest and most effective medicines needed in a health system.

ASA is a non-steroid anti-inflammatory drug with anti-platelet properties that works primarily through inhibition of COX-1 and COX-2. COX-1 regulates the production of prostacyclin and TXA

2

. Prostacyclin is a potent vasodilator and inhibitor of platelet aggregation, while TXA

2

is a potent vasoconstrictor and promotes platelet aggregation.

A low dose of ASA has been used during pregnancy mostly to prevent the onset of preeclampsia. Other suggested indications have included prevention of RPL, preterm birth and fetal growth restriction. The majority of RCTs have found no increase in hemorrhagic complications, placental abruption, postpartum hemorrhage or mean blood loss associated with low-dose ASA during pregnancy (ACOG Committee Opinion, 2018).

When it comes to the fetus, no risk of congenital anomalies has been found, although

a possible, but very weak, association between ASA use during pregnancy and

gastroschisis has been pointed out (Werler et al., 2002; Kozer et al., 2002).

(33)

33

Administration of ASA (60-150 mg) in the third trimester has not been associated with ductal closure (Sibai et al., 1989), persistent pulmonary hypertension, neonatal intracranial hemorrhage or other neonatal hemorrhagic complications (Duley et al., 2007). In a study by Werler, the hypothesis that ASA taken during first-trimester pregnancy would increase the risk of congenital heart defects was tested. No increased risk, compared to that of other structural malformations, was found (Werler et al., 1989). The question of neonatal malformations related to ASA 50-150 mg daily during pregnancy was evaluated, based on Swedish Medical Birth Register data, by the Pharmacological Council in Region Stockholm, Sweden, in 2018. The studied group comprised 6,601 children whose mothers had used ASA during pregnancy to inhibit platelet aggregation. The total rate of malformations was normal, 2.3% compared with the expected 2.1%. An increased rate of cardiac malformations was reported in the ASA group: 63 compared with the expected 46.

Twenty-nine had a ventricular septum defect, but 21 had an atrial septum defect, compared with the expected nine. The rate of hydronephrosis was also increased: 19, compared with the expected nine. The authors did not make any recommendation concerning ASA prescription during pregnancy.

When it comes to the effect of anticoagulant treatment in reducing miscarriage rates or improving LBR, studies report diverging results. The interventions in the following briefly presented studies consist of ASA alone, LMWH alone, or ASA and LMWH in combination (Table 4).

Only one RCT was found comparing low-dose ASA with placebo (Tulppala et al., 1997). The other trials compared ASA with other drugs, alone or in combination.

There was also a variation in study design.

Based on a meta-analysis (de Jong, 2014) and the results of two subsequent large

RCTs (Pasquier et al., 2015; Schleussner et al., 2015), there was no evidence that

LMWH alone, ASA alone or the two in combination improved the LBR in women

with consecutive unexplained RPL (ESHRE Guideline Nov., 2017).

(34)

34

Table 4. Interventional studies of anti-platelet and anti-coagulant treatments in women with unexplained recurrent pregnancy loss

ASA=acetylsalicylic acid, gw=gestational week, LBR=live birth rate, RCT=randomized controlled trial;

TLC=tender loving care Author,

year

Study design, participants, n

Previous miscarriages

Treatment, participants, n

Outcome Comments

Pasquier 2015

Multicenter RCT n=256

≥2 consecutive before gw 15

Enoxaparin 40 mg (n=138) / placebo (n=118)

Live birth 66.6% vs. 72.9%

Stratified for no.

of previous miscarriages and age

Schleussner 2015

Multicenter RCT n=434

Dalteparin 5000 IU (n=220)/

multivitamin pills (n=214)

Ongoing pregnancy at gw 24

86.8% vs. 87.9%

Clark 2010

Multicenter RCT (SPIN) n=294

Enoxaparin 40 mg and ASA 75 mg and TLC (n=147)/

TLC alone (n=147)

Miscarriage 22% vs. 20%

Kaandorp 2010

Multicenter RCT (ALIFE) n=364 (enrolled), n=310 (pregnant)

≥2 before gw 20 ASA 80 mg and fraxiparine 2850 IU (n=103)/ASA 80 mg alone (n=104)/placebo (n=103)

Live birth 54.5%, 50.8% and 57.0%

400 µg folic acid to all participants during pregnancy

Badawy 2008

Single-center RCT

n=340

≥3 first-trimester Enoxaparin 20 mg (n=170)/no treatment (n=170)

Miscarriage 4.1% vs. 8.8%

(early) 1.1% vs.

2.3% (late)

500 µg folic acid until gw 13 to all participants

Dolitzky 2006

Multicenter RCT, single-blind n=107

≥3 consecutive first-trimester or

≥2 second- trimester

Enoxaparin 40 mg (n=54)/ASA 100 mg (n=50) Lost to follow-up (n=3)

No placebo group

Tulppala 1997

Single-center RCT

n=54

ASA 50 mg (n=27)/placebo (n=27)

Elmahashi 2014

Single-center RCT, open n=150

≥3 consecutive Enoxaparin 40 mg and ASA 75 mg (n=75)/ASA 75 mg (n=75)

Miscarriage 29% vs. 47%

Live birth 71% vs. 42%

High miscarriage rate, low LBR and preterm delivery rate about 23%

Rai 2000

Single-center observational study n=805

ASA 75 mg (n=367)/no treatment (n=438)

(35)

35

Immunoglobulin

Jablonowska conducted a double-blind RCT (n=41) to evaluate the effect of treatment with IVIG (20g/400ml Gammonativ, n=22), compared with placebo (400ml saline, n=19), in order to prevent new miscarriages in a group of women with three or more consecutive unexplained miscarriages before gw 20. All previous pregnancies were documented by ultrasound or histology. The LBR was 77% for IVIG and 79% for placebo (Jablonowska et al., 1999). Another five RCTs were evaluated by Rasmark Roepke in a systematic review. IVIG was compared with placebo in four studies, but the placebo drugs were different (three albumin and one saline). IVIG was compared with ASA+LMWH in the fifth study. No significant effect on LBR was shown (Rasmark Roepke et al., 2018).

Leucocyte immunization

Gatenby published a double-blind RCT comparing immunization with 400 million paternal lymphocytes and immunization with 400 million autologous lymphocytes, in women with at least three consecutive confirmed miscarriages before gw 20.

Treatment started before pregnancy and was repeated every 12 weeks until positive pregnancy test. LBRs were 68% and 47% for paternal and autologous lymphocyte immunization, respectively (p=0.1) (Gatenby et al., 1993). In the systematic review by Rasmark Roepke, another four RCTs were evaluated. There were design differences between all trials in terms of start and of testing time-points. The trials were old (published between 1985 and 1994) the results were not trustworthy due to very low certainty of evidence based on a high risk of bias and severe imprecision (Rasmark Roepke et al., 2018).

Psychological support/tender loving care

As early as in in 1954, Javert wrote about the need for psychotherapy for patients with RPL. A therapeutic regimen, including a pre-conceptional consultation for the patient and her partner, and examination for detection of specific medical, gynecological and psychological factors, was offered. Frequent prenatal visits and unlimited phone calls in order to relieve anxiety were also included. “It has been our practice since 1940 to establish a pre-conceptional interview with the patient. The patient was allowed to verbalize her past experiences with the miscarriages, doctors, hospitals and her attitude toward another pregnancy. Then she was referred, with her partner, to the psychosomatic consultant for more detailed review” (Javert, 1954).

When these patients had had a successful pregnancy, their subsequent pregnancies

were uncomplicated. According to this author, repeat miscarriage is in itself a stress

mechanism which may lead to rejection of the next pregnancy (Javert, 1954).

(36)

36

Musters did something very similar: a questionnaire study of 174 women with at least two previous miscarriages. The women preferred a plan for the first trimester in a subsequent pregnancy that involved one single doctor and repeated ultrasounds, as well being shown understanding and being listened to. Furthermore, they wanted healthcare staff to be aware of their obstetric history and show respect for them and their miscarriages (Musters et al., 2013).

Couples with recurrent miscarriage have higher levels of depression, anxiety and feelings of guilt, compared with the general population. Therefore, patients may benefit from regular contact with a dedicated miscarriage clinic (Morley et al., 2013;

Fertl et al., 2009; Homer, 2019; Rai et al., 2006). When a follow-up clinic was established, 79% of women with previous recurrent miscarriage attended, and they all found the contact helpful (Lee et al., 1996).

A third of the women seeing psychiatric specialists due to RPL are clinically depressed, and about 20% have anxiety symptoms similar to those in general psychiatric outpatient populations. It is evident that anxiety symptoms may occur in response to miscarriage (Lee et al., 1996).

In 1962, Tupper and Weil suggested that psychotherapy should be included as part of the treatment of RPL. A group of women with three previous consecutive miscarriages, given psychotherapy in the subsequent pregnancy, had a LBR of 84%, compared to 26% in the control group (Tupper et al., 1962). James also concluded that psychotherapy had a positive effect in reducing RPL, as the LBR was 80% in the therapy group, compared with 42% in the group not given therapy (James et al., 1963). However, the small study populations and a major risk of selection bias, constitute limitations in these three old observational studies. The results are thus not applicable.

Supportive care as the sole treatment has been described in several case series, reporting 70-86% LBR in subsequent pregnancies in women with unexplained RPL (Badawy et al., 2008; Fawzy et al., 2008; Brigham et al., 1999; Bricker et al., 2002).

In a study by Stray-Pedersen of couples with three or more unexplained consecutive miscarriages, the LBR was 86% after special psychological support, compared with 33% after routine prenatal care (Stray-Pedersen et al., 1984). The same levels, 86%

and 33%, respectively, were found by Liddell, although the fact that the control group

included only nine women was a limitation (Liddell et al., 1991). These authors

(37)

37

concluded that psychological support alone may lead to a significant improvement in the outcome of pregnancy.

In summary, the studies on supportive care alone are all observational and suffer

from major methodological drawbacks. Rai suggests that patients with RPL should

be referred to placebo-controlled trials, in order to elucidate the value, shown in

previous studies, of psychological support in improving pregnancy outcome (Rai et

al., 2006). This suggestion is wise, since all individuals participating in prospective

trials will be thoroughly monitored and taken care of, which will likely be beneficial

for women with RPL.

(38)

38

(39)

39

Aims

Aims of the thesis;

 to investigate whether a low dose of ASA, 75 mg daily, during pregnancy can prevent a new miscarriage in women with at least three consecutive unexplained first-trimester miscarriages in the current relationship (Paper I)

 to study platelet aggregation in healthy women during normal pregnancy longitudinally by multiple electrode impedance aggregometry and compare it with platelet aggregation in the non-pregnant state (Paper II)

 to compare platelet aggregation during pregnancy in women with and without RPL for unknown reasons (Paper III)

 to study the effect of ASA on platelet aggregation in women with RPL during pregnancy longitudinally and in comparison with healthy pregnant women (Paper III)

 to investigate the effect of TPO-ab and of TSH levels in the upper normal

reference range on the risk of first-trimester miscarriage in women with a

history of early RPL (Paper IV)

(40)

40

(41)

41

Methods

Study populations

The main study population consisted of women with RPL. One RCT (Paper I) and two post hoc studies (Papers III and IV) were derived from this screened population.

A population of healthy women with normal pregnancy constituted the study population in Paper II and contributed a comparison group in Paper III. The formation of the final study populations in the four papers is presented in two flow charts (Figures 8 and 9). The specific methodology related to each paper is presented below.

Fig. 8. Flow chart of the study populations of women with recurrent early pregnancy loss.

Abbreviation: ASA-RCT=randomized controlled trial on preventive effect of acetylsalicylic acid for recurrent miscarriage; TPO-ab=thyroid peroxidase antibodies

(42)

42

Fig. 9. Flow chart of the study populations of healthy women with normal pregnancy.

Abbreviation: AA=arachidonic acid

Recurrent unexplained first-trimester miscarriage